225 225225 225225 Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 94(2): 225-228, Mar./Apr. 1999
RESEARCH NOTE
The Evolution of Salivarian
Trypanosomes
Jamie Stevens
+, Wendy Gibson
School of Biological Sciences, University of Bristol, Bristol BS8 1UG, UK
Key words: Salivaria Trypanosoma brucei Trypanosoma congolense Trypanosoma rangeli
-evolution - 18S ribosomal RNA
CA Hoare (1972 The Trypanosomes of
Mam-mals, Blackwell, Oxford, 749 pp.) divided
mam-malian trypanosomes into two sections, the Salivaria and the Stercoraria. The Salivaria com-prises all the African tsetse-transmitted trypano-somes of mammals, while the Stercoraria includes mammalian trypanosomes transmitted by the pos-terior route. However, the status of other tsetse-transmitted species, e.g. Trypanosoma grayi from crocodiles, and of other trypanosomes transmitted by the anterior route, e.g. T. rangeli from South America and leech transmitted trypanosomes, re-main under debate. To some extent such unknowns are linked to the limitations of the morphological and transmission characters used in early investi-gations of trypanosome taxonomy and evolution. With the wealth of molecular sequence data now available, particularly for the phylogenetically ver-satile 18S rRNA gene (ML Sogin et al. 1986 Proc
Natl Acad Sci USA 83: 1383-1387) some of these
questions may now be fruitfully re-addressed. Accordingly, we undertook phylogenetic analysis of published ribosomal RNA 18S sequences, supplemented with data from our own studies, our aim being to re-examine the evolutionary relation-ships of Salivarian trypanosomes of mammals and other vertebrates.
Fourteen Trypanosoma species 18S rRNA se-quences were obtained from EMBL/GenBank for phylogenetic analysis: T. boissoni U39580; T.
carassii L14841; T. rotatorium U39583; T. triglae
U39584; T. brucei brucei M12676; T. congolense (kilifi-type) U22317; T. congolense (forest-type) U22319; T. congolense (savannah-type) U22315;
T. simiae U22320; T. vivax U22316; T. cruzi
X53917; T. cruzi M31432; T. avium U39578; T.
scelopori U67182; together with three newly
avail-able sequences (JR Stevens et al. 1999
Parasitol-ogy 118: 107-116.); T. grayi AJ005278; T. rangeli
AJ009160; T. varani AJ005279. Five additional 18S sequences from a range of other kinetoplastid species were included as outgroups: Crithidia
fasciculata X03450; Leishmania amazonensis
X53912; L. major X53915; Trypanoplasma borreli L14840; Bodo caudatus X53910. Sequences were aligned primarily on the basis of their secondary structure (J-M Neefs et al. 1990 Nucleic Acids Res
18: 2237-2243). Sub-sections of the alignment,
between regions of high homology were sub-aligned using the program Clustal V (DG Higgins et al. 1992 Comp Applns Biosci 8: 189-191), be-fore final adjustments were made by eye. Bootstrapped maximum parsimony analysis of the 22 rRNA 18S sequences was performed with 100 replicates using test version 4.0d63 of PAUP*, written by David L Swofford.
The phylogenetic analysis (Figure) places the Salivarian trypanosomes in a monophyletic clade comprising exclusively mammalian trypanosomes of African origin. Within the Salivarian group, the various types of T. congolense also constitute a monophyletic group. T. rangeli, T. grayi and try-panosomes of fish and amphibia are excluded from the Salivarian clade; T. rangeli is placed firmly in a separate clade with T. cruzi. A third major clade, comprising trypanosomes with aquatic hosts, forms a separate early branch within the monophyletic
Trypanosoma which is not directly ancestral to
ei-ther the Salivaria or the Stercoraria. Branches re-ceiving less than 50% bootstrap support are pre-sented as polytomies.
In agreement with a number of previous stud-ies (F Alvarez et al. 1996 Mol Phylogen Evol 5: 333-343, J Lukes et al. 1997 J Mol Evol 44: 521-527), our analysis confirms the monophyly of the African Salivarian trypanosomes. Within the Salivaria, our results also indicate that the various types of T. congolense (forest, kilifi, savannah) share common ancestry. Such a finding contrasts with the results of isoenzyme and RAPD based studies (I Sidibé, cited M Tibayrenc 1998 Int J
Parasitol 28: 85-104), in which it is suggested that T. congolense may be polyphyletic. We propose
that this apparent difference in conclusions is due
This work was financed by The Wellcome Trust (Grant No 047131/Z/96/Z).
+Corresponding author. Present address: Hatherly
Labo-ratories, Department of Biological Sciences, University of Exeter, Prince of Wales Road, Exeter EX4 4PS, UK. Fax: +44-1392-263700. E-mail: [email protected] Received 12 November 1998
226 226 226 226
226 The Evolution of Salivarian Trypanosomes J Stevens, W Gibson
to the differing clock speeds of the markers used in the two studies, highlighting the potential limi-tations of using fast evolving ‘population genet-ics’ markers for evolutionary studies.
The taxonomic and evolutionary status of T.
rangeli has long been debated (A D’Alessandro &
NG Saravia 1992 Trypanosoma rangeli, p. 1-54. In JP Kreier & JR Baker (eds), Parasitic
Proto-zoa, 2nd ed., vol. 2, Academic Press, San Diego)
and, even following the application of molecular tools (for example, MI Amorim et al. 1993 Acta
Tropica 53: 99-105), has remained unresolved. The
results of this study, however, indicate a close evo-lutionary relationship between T. rangeli and T.
cruzi, suggesting T. rangeli be placed within the
Stercorarian subgenus Schizotrypanum. Similarly, the non-Salivarian status of the posteriorly tsetse-transmitted T. grayi is also confirmed; the
evolu-tionary significance of its position separate from either the Salivaria or the Schizotrypanum species (i.e. T. cruzi and T. rangeli) remains to be explored. Finally, the distinct nature of the clade com-prising trypanosomes with aquatic hosts, which forms a separate early branch within the monophyl-etic Trypanosoma, suggests that these taxa are not directly ancestral to either the Salivaria or the Stercoraria. The position of this ‘aquatic’ clade may have fundamental consequences for hypoth-eses concerning the evolution of the Salivaria and parasitism in the Kinetoplastida as a whole.
DISCUSSIONS
Dear Drs Stevens and Gibson
Thank you for your very interesting presenta-tion. I found your results very interesting but I would also like to ask two further questions: (i)
Phylogram constructed by bootstrapped (100 replicates) maximum parsimony analysis of 22 kinetoplastid 18S ssu rRNA se-quences, based on a standard alignment of 1809 nucleotide sites (JR Stevens et al. 1999 Parasitology 118: 107-116). Bootstrap values for all major nodes are given and all branches receiving bootstrap support values >50% are shown; relationships failing to achieve this level of support are shown as polytomies.
227 227227 227227 Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 94(2), Mar./Apr. 1999
The level of bootstrap support for your different clades. I did not find this important information at your website either; (ii) What strain you used for
T.rangeli. This is a polymorphic species and its
strong association with T. cruzi (another highly polymorphic species) in your study could be strain dependant.
Hooman Momen
Answer (i): apologies for not giving the bootstrap values, this was a serious oversight. Support for the major clades were as follows: rangeli/cruzi clade >91%; Salivarian clade 98%; aquatic clade >91%. I hope this goes some way to convincing you of the robustness (and our associated enthusi-asm) of these results.
Answer (ii): T. rangeli strain RGB(Basel). Per-haps some of you could furnish us with more his-tory and results relevant to this strain.
Dear Drs Stevens and Gibson
Did you investigate alternative topologies from the point of view of how much “worse” they are? Thank you for your presentation.
Dmitri Maslov
University of California - Riverside, Department of Biology 3401 Watkins Drive, Riverside, CA 92521, USA E-mail: [email protected]
Answer: I have begun to explore the tree with the various branch swopping tools available in MacClade and, thus far, all the major clades (Salivarian, Schizo and aquatic) are robust. The early branching of the aquatic clade will require further investigation, but, whatever our findings, it will not affect significantly our conclusions re-lating to the relationship of the Salivarian and Schizo clades. I hope to explore clade robustness further in the near future using Bremer support measures.
Dear Jamie and Wendy
I was very interested by your phylogenetic study. However, I was surprised that you conclude
T. rangeli is a Schizotrypanum from the fact that it
is closer to T. cruzi than to African trypanosomes. Hoare (1972) classified other species as stercorarian, such as T. lewisi, T. musculi (subge-nus Herpetosoma) and T. theileri and T.
melophagium (subgenus Megatrypanum). Couldn’t
it be that T. rangeli is closer to these species than to T. cruzi ? Or do you have evidence that T. rangeli is phylogenetically closer to T. cruzi than are other Schizotrypanum species such as T. dionisii or T.
vespertilionis?
Sylvain Brisse
Answer: as noted above, support for the close re-lationship of T. rangeli and T.cruzi (based on this 18S marker) is good (>91%) and suggests a genu-ine phylogenetic relationship with T.cruzi. Whether or not T. rangeli is phylogenetically closer to T.
cruzi than are some other Schizotrypanum
spe-cies remains to be seen, but even if it is not, it does not rule against its Schizotrypanum status ... in any phylogenetic or taxonomic system, some taxa will be obviously be more or less integral to the clade.
COMMENTS
The taxonomic position of T. rangeli has been a point of contention for many years. However, this species can be distinguished from members of the subgenus Schizotrypanum by a number of methods (reviewed by Grisard et al). It is well known that these two species occur in the same hosts. There is the famous case when a mixed infection of the two species was described as a new species. We must be open to new data but we need to be certain that the DNA that is being analysed is from the correct para-site. What other data is available to say that the strain used by Drs Stevens and Gibson was T. rangeli? It would be interesting to analyse the actual strain fur-ther with ofur-ther methods.
We must remember what emphasis was popu-lar in biology at the time that a particupopu-lar classifi-cation system was proposed. When Dr Hoare made his division of the trypanosomes into hind and forgut transmission biological characters were begining to gain ground against classical morphol-ogy. Such characters are indeed a summation of many different complex characters which may or may not be homologous. Thus although a devel-opment site is a definitely may be useful for char-acterizing an organisms it may or may not be use-ful in classification depending on its homology within the group. My point is that the character stercoraria or salivaria is useful for describing where the infective forms are found or leave the vector but we must be careful in using it in classi-fication. Thus to me it is quite acceptable that T.
rangeli and T. brucei have a salivarian
develop-ment since the infective forms are found in the mouthparsts. However, that does not mean that we have to classify them in the same group. I hope this goes some way to answering Roberto’s ques-tion and clarrifying the use of the terms salviarian and stercorarian.
Jeffrey Shaw
It is well known that T. rangeli and T. cruzi share a large number of triatomine vectors and vertebrate hosts, including man. For T. cruzi, many studies describing such type of variables have been published. Until the finding of T. rangeli in
south-228 228 228 228
228 The Evolution of Salivarian Trypanosomes J Stevens, W Gibson
ern Brazil, the distribution area of this parasite was restricted in Brazil to the Amazon region. Unfor-tunately, I don’t know specific papers describing (or identifying) variables for such approach, prob-ably due to the non-pathogenic characteristic of T.
rangeli to humans. Among several characteristics,
both of these human trypanosomes are transmitted by triatomines, however, through different ways. Thus, the expected distribution for T. rangeli is correspondent to the distribution of susceptible triatomine vectors. These triatomines in which T.
rangeli can invade and develop in the hemolymph
and salivary glands and are transmitted by bite mainly belongs to the Genus Rhodnius. As you see, this expected distribution for T. rangeli is overllaping the T. cruzi distribution area. I agree with your comment that studies targeting variables to predict the distribution of T. rangeli, or the pos-sible spreading area of this parasite, will be very helpful and useful to better understand the biology of this parasite.
Concerning the discussions and questions about
T. rangeli taxonomic position: Dr Hoare classified
trypanosomes into Salivaria and Stercoraria sec-tions based in their development mode, primarily in the vector and secondarily in the mammalian hosts. In his monograph published in 1972, he de-fined as Stercoraria those trypanosomes which development in the vector is completed (with for-mation of metatrypanosomes) in the posterior sta-tion and transmission is contaminative (in T. rangeli also in anterior station, with inoculative transmis-sion). He defined as Salivaria those trypanosomes which development in the vector is completed (with formation of metatrypanosomes) in the anterior station (except in mechanical inoculators) and transmission is inoculative. There are some remark-able facts that we have to keep in mind since Dr Hoare initial description. First, there is no descrip-tion of successful transmission of T. rangeli through forms present in triatomine feces, despite the pres-ence of this parasite in both naturally and experi-mentally infected triatomines. T. rangeli metacyclic trypanosomes only develop inside salivary glands of triatomine bugs and are transmitted to the ver-tebrate host through inoculation with saliva dur-ing blood suckdur-ing. For these reasons, Dr Hoare mentioned T. rangeli as an exception in the Stercoraria section, even sharing some other bio-logical characteristics with the type species (T.
cruzi). As very well noted by Dr Shaw, we think
that this nomenclature was based on biological characteristics and focused primarily on the site of development of the infective forms inside the
vec-tor. More studies should be addressed to better understand the particular evolutionary correlation of these different transmission pathways. The use of modern biochemical, immunological and mo-lecular methods for taxonomic purposes of trypanosomatids should be treated with special care. The number of strains, their biological be-havior, isolation methods, geographical regions and original fonts as well as the high polymorphism detected among strains or stabilates within same species should be considered in such analysis. Also, relying on single or even a few methods or mark-ers can result in dubious categorizations due to the narrow biological point of view of these analyses. Phylogenies intended to resolve the classification of these closely related organisms should be based on markers more relevant to the processes of trans-mission. Trypanosome strains should be very well characterized prior to analysis using such refined methods, since contaminations might be expected, especially in labs that cultivate both parasites (T.
cruzi/T. rangeli) or in regions were these parasites
coexists. Recently, Dr H Noyes (Parasitol Today
14: 49-50, 1998) reviewed the credibility of
trypa-nosome trees. Different mechanisms that can af-fect and/or distort these trees, giving rise to mis-interpretations are discussed in this very interest-ing paper. Despite T. rangeli sharinterest-ing characteris-tics with different subgenus of both Salivaria and Stercoraria sections, including the 18s SSU rRNA sequence described in this meeting, it is transmit-ted through triatomine bite, i.e. through anterior station. In reference to other points raised in rela-tion to the paper of Stevens and Gibson, T. rangeli forms present in triatomine faeces are, at least un-til now, non-infective for vertebrate hosts. Thus, due to this biological characteristic we believe that
T. rangeli is not properly classified in the
Stercoraria section. Moreover, the suggestion to treat T. rangeli as a Schizotrypanum trypanosome due to the close relationship with T. cruzi revealed only by SSU rRNA analysis of a single strain is too dangerous. Our aim was not to propose the change of T. rangeli taxonomic position, but to evaluate the heterogeneity among different isolates and bring up a very old a controversial question that remains unanswered. We probably should use as many well characterized representative strains and characteristics as possible to make evolution-ary or phylogenetic inferences.
Edmundo Grisard
Acknowlegments: to F Tarrieu and M Tibayrenc, CEPM, Monpellier, for supplying T. rangeli genomic DNA.
